With the increasing frequency of precipitation events under global warming, understanding rainfall-induced disruptions to urban mobility has become increasingly important. While prior studies primarily focus on road traffic, the lagged and threshold effects of rainfall on urban rail transit (URT) passenger flow remain insufficiently explored. This study analyzes 109 days of automatic fare collection data from Tianhe District, Guangzhou, in combination with hourly meteorological records and station-level built environment attributes. A rainfall threshold-aware gradient boosting framework is proposed to capture nonlinear response regimes, and an explainable learning approach is used to quantify the relative importance of rainfall, temporal factors, and built environment characteristics. The proposed framework outperforms the baseline model, with the root mean squared error (RMSE) and mean absolute error (MAE) reduced by over 5.38% and 5.93%, respectively. Results further indicate that lagged rainfall intensity exerts the strongest influence on passenger flow variation, with impact magnitudes varying systematically across station types. These findings enhance understanding of the nonlinear, time-dependent effects of rainfall on URT demand and provide practical guidance for passenger flow management and operational planning under rainfall conditions.
Li et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: